The human milk microbiome changes over lactation and is shaped by

maternal weight and mode of delivery14

Raul Cabrera-Rubio, M Carmen Collado, Kirsi Laitinen, Seppo Salminen, Erika Isolauri, and Alex Mira

INTRODUCTION

Breast milk gained its nutritional superiority after the divergence of mammals millions of years ago (1). Breast milk
contains several functional and protective nutrients, which help to
create the right microenvironment for gut development and
maturation (2, 3). Protection is provided by components including regulatory cytokines and growth factors. In addition,
breast milk contains several other factors such as lysozyme,
lactoferrin, and oligosaccharides, which assist in preventing
infections and supporting the growth of beneficial bacteria.

544

Furthermore, breast milk is also a continuous source of microbes,

their growth factors, and components that regulate host-microbe
interactions. These factors emphasize the key position of breastfeeding in conferring protection during a critical period in life,
when breast milk is the sole source of nutrition for the infant and
when the neonates own immune defenses, including the integrity
of the gut barrier, are immature (47).
Breastfeeding has been reported to modify infant gut microbiota development (810), and breast milk is recognized as one
of the most important postpartum elements modulating metabolic and immunologic programming related to the childs
health (11). In overweight and obese pregnant women, a vicious
circle of unfavorable metabolic development may be generated
if the aberrant gut microbiota associated with overweight and
obese or excessive weight gain during pregnancy is transferred
to the infant (1215). Differences in microbiota composition in
infants delivered by cesarean delivery compared with vaginally
born infants (15, 16) suggest that mother-infant transmission of
microbiota occurs during vaginal delivery. Recent work has
opened up new angles in the uncovering of a previously unknown source of microbiota, that in breast milk (17), which
suggests that maternal factors can influence the composition and
activity of the microbiota. The purpose of the present study was
thus to identify pre- and postnatal factors that can potentially
influence the bacterial communities inhabiting human milk and
to compare the bacterial composition of breast milk with that
1
From the Joint Unit of Research in Genomics and Health, Centre for
Public Health Research (CSISP)Cavanilles Institute for Biodiversity and
Evolutionary Biology, University of Valencia, Valencia, Spain (RC-R and
AM); the Institute of Agrochemistry and Food Science, Spanish National
Research Council (IATA-CSIC), Paterna, Valencia, Spain (MCC); Functional Foods Forum (KL and SS) and the Institute of Biomedicine (KL),
University of Turku, Turku, Finland; and the Department of Pediatrics, Turku University Hospital and University of Turku, Turku, Finland (EI).
2
RC-R and MCC contributed equally to this article.
3
Supported by Fun-C-Food CSD2007-00063, Microgen CSD2009-00006
from the Consolider-Ingenio program, SAF2009-13032-C02-02 from the
Spanish Ministry of Science and Innovation, and grants 069/2010 and 071/
2011 from Consellera de Sanidad, Generalitat Valenciana, Spain, and from
Academy of Finland (project no. 2600006911). MCC is the recipient of
a Ramon y Cajal research contract (RYC-2010-05614) from the Spanish
Ministry of Science and Innovation.
4
Address correspondence to MC Collado, IATA-CSIC, Av Agustin Escardino 7, 49860, Paterna, Valencia, Spain. E-mail: mcolam@iata.csic.es.
Received February 21, 2012. Accepted for publication May 30, 2012.
First published online July 25, 2012; doi: 10.3945/ajcn.112.037382.

Am J Clin Nutr 2012;96:54451. Printed in USA. 2012 American Society for Nutrition
Supplemental Material can be found at:
http://ajcn.nutrition.org/content/suppl/2012/08/21/ajcn.112.0
37382.DC1.html

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

ABSTRACT
Background: Breast milk is recognized as the most important postpartum element in metabolic and immunologic programming of
health of neonates. The factors influencing the milk microbiome
and the potential impact of microbes on infant health have not yet
been uncovered.
Objective: Our objective was to identify pre- and postnatal factors
that can potentially influence the bacterial communities inhabiting
human milk.
Design: We characterized the milk microbial community at 3 different time points by pyrosequencing and quantitative polymerase
chain reaction in mothers (n = 18) who varied in BMI, weight gain,
and mode of delivery.
Results: We found that the human milk microbiome changes over
lactation. Weisella, Leuconostoc, Staphylococcus, Streptococcus,
and Lactococcus were predominant in colostrum samples, whereas
in 1- and 6-mo milk samples the typical inhabitants of the oral
cavity (eg, Veillonella, Leptotrichia, and Prevotella) increased significantly. Milk from obese mothers tended to contain a different
and less diverse bacterial community compared with milk from normalweight mothers. Milk samples from elective but not from nonelective
mothers who underwent cesarean delivery contained a different bacterial community than did milk samples from individuals giving
birth by vaginal delivery, suggesting that it is not the operation
per se but rather the absence of physiological stress or hormonal
signals that could influence the microbial transmission process to
milk.
Conclusions: Our results indicate that milk bacteria are not contaminants and suggest that the milk microbiome is influenced by
several factors that significantly skew its composition. Because bacteria present in breast milk are among the very first microbes entering the human body, our data emphasize the necessity to understand
the biological role that the milk microbiome could potentially play
for human health.
Am J Clin Nutr 2012;96:54451.

545

FACTORS INFLUENCING HUMAN MILK MICROBIOME

from other sites in the human body. For this purpose, we analyzed breast-milk composition through the recommended period
of breastfeeding in women who varied in BMI, weight gain
during pregnancy, and mode of delivery.
SUBJECTS AND METHODS

Subjects and design

Breast-milk samples
Breast-milk samples from mothers were collected within 2 d
after mothers gave birth in the maternity hospital (colostrum) and
at 1 and 6 mo after delivery at home. Mothers were given written
instructions for standardized collection of samples each morning,
and the samples were frozen and stored at 2208 C for later
analysis. Before sample collection, the breast was cleaned with
an iodine swab to reduce bacteria residing on the skin, and
breast milk was collected manually, discarding the first drops,
with a sterile milk collection unit.
DNA extraction and quantitative real-time polymerase
chain reaction analysis
Before analysis, the milk samples corresponding to colostrum
and 1 and 6 mo of infant age were thawed and centrifuged at
10,000 3 g for 10 min to separate cells and fat from whey.
Thereafter, total DNA was isolated from the pellets by using the
QIAamp DNA Stool Mini Kit (Qiagen). Quantitative polymerase chain reaction (qPCR)5 and specific bacterial group
analyses were conducted as previously described (12, 13). qPCR
amplification and detection were performed with an ABI PRISM

5
Abbreviations used: OTU, operational taxonomic unit; PCA, principal
components analysis; PCR, polymerase chain reaction; qPCR, quantitative
polymerase chain reaction.

Values
Maternal age (y)
Weight before pregnancy (kg)
Normal weight
Obese
BMI before pregnancy (kg/m2)
Normal weight
Obese
Maternal weight gain over pregnancy (kg)
Normal weight
Obese
Duration of pregnancy (wk)
Method of delivery (% vaginal)
Medical induction of labor by using Cytotec
(Pfizer; oxitocin) (%)
Problems during pregnancy (%)
Normal, no problems
Gestational diabetes
Streptococcus B positive
Other infections
Parity (%)
1
2
.2
Sex (% female)
Infant head circumference at birth (cm)
Infant weight (kg)
At birth
Age 6 mo
Age 12 mo
Age 24 mo
Infant length (cm)
At birth
Age 6 mo
Age 12 mo
Age 24 mo
Maternal antibiotic treatment (%)
Before delivery
During delivery
After delivery
Infant antibiotic treatment (%)
At birth
First months
Total time exclusively breastfeeding (mo)
Breastfeeding at 1 mo (%)
Breastfeeding at 6 mo (%)
Infant diet at age 6 mo (%)
Exclusively breast milk
Breast milk + formula
Breast milk + solid foods
Breast milk + formula + solid foods
Formula

32.01
76.11
62.88
86.70
27.71
22.90
31.86
12.45
14.70
10.49
40.40
50.0
11.1

6 5.121
6 13.87
6 4.58
6 8.17
6 4.77
6 1.11
6 2.22
6 5.82
6 5.60
6 4.60
6 1.11
(9/18)2
(2/18)

88.8
11.1
0
0

(16/18)
(2/18)
(0/18)
(0/18)

33.3
33.3
33.3
55.5
35.28

(6/18)
(6/18)
(6/18)
(10/18)
6 1.11

3.69
8.13
9.87
12.41

6
6
6
6

0.32
0.90
1.40
1.03

51.25
68.80
76.50
87.67

6
6
6
6

1.84
2.82
3.65
2.80

0 (0/18)
16.6 (3/18)
5.5 (1/18)
0
0
4.10
100
100

(0/18)
(0/18)
6 1.10
(18/18)
(18/18)

0
0
66.7
33.3
0

(0/18)
(0/18)
(12/18)
(6/18)
(0/18)

Mean 6 SD (all such values).

Percentage and prevalence reflect the percentage of the number of
positive out of total samples included in the study.
1
2

7300-PCR sequence detection system (Applied Biosystems).

Each reaction mixture of 25 mL was composed of SYBR Green
PCR Master Mix (Applied Biosystems), 0.5 mL of each of the
specific primers at a concentration of 0.25 mmol/L, and 1 mL
template DNA. The fluorescent products were detected in the
last step of each cycle. A melting curve analysis was performed

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

Mothers were selected from an ongoing prospective, randomized follow-up study that started in April 2002 (NCT00167700,
section 3; registered through clinicaltrials.gov) (18). Pregnant
women were recruited at their first visit to a maternal welfare
clinic if they had no metabolic or chronic diseases other than
allergy. For the present report women were selected according to
their prepregnancy BMI, early exclusive breastfeeding practices,
and breastfeeding follow-up during 6 mo as well as breast-milk
sample availability. A total of 18 women were included: 10 obese
women with a BMI (in kg/m2) .30 and 8 normal-weight women
with a BMI #25 who served as controls. Details of delivery and
early infant feeding were collected after birth. Data on duration
of breastfeeding as well as infant feeding practices were obtained by interview and recorded at each visit. Maternal dietary
guidance was provided by a nutritionist during each study visit
to conform to the current dietary recommendations. Written
informed consent was obtained from the participants, and the
Ethics Committee of the Hospital District of South-West Finland
approved the study protocol. Normal weight gain was defined
according to the Institute of Medicines recommendations: 11.5
16.0 kg and 7.011.5 kg for normal-weight and obese women,
respectively. Clinical characteristics of mother-infant pairs are
shown in Table 1.

TABLE 1
Clinical characteristics of the infants and their mothers (n = 18 motherinfant pairs) included in the study

546

CABRERA-RUBIO ET AL

after amplification to distinguish the targeted from the nontargeted polymerase chain reaction (PCR) product. The bacterial
concentration in each sample was calculated by comparing the
Ct values obtained from standard curves. These were created by
using serial 10-fold dilution of pure culture-specific DNA fragments corresponding to 10109 gene copies/mL. Statistical analyses were performed by using SAS version 9.2 software for
Windows (SAS Institute Inc). For the BMI and weight gain
groups, microbial data were expressed as medians with IQRs, and
the Mann-Whitney test was used for comparisons. Differences in
the prevalence of bacterial groups were established applying the
chi-square test.

PCR amplification and pyrosequencing

Sequence analysis
Sequences with an average quality value of <20 and/or with
.4 ambiguities in homopolymeric regions in the first 360 flows
were excluded from the analysis. Only sequences longer than
200 bp were considered. Sequences were assigned to each
sample by the 8-bp barcode and passed through the Ribosomal
Database Project classifier software (21), where each sequence
was assigned a phylum, class, family, and genus, as long as the
taxonomic assignment was unambiguous within an 80% confidence threshold. To estimate total diversity, sequences were
clustered at 97%, 95%, and 90% nucleotide identity over a 90%
sequence alignment length by using the Ribosomal Database
Project pyrosequencing pipeline. For this analysis, sequences
.97% identical were considered to correspond to the same
operational taxonomic unit (OTU), representing a group of sequences that presumably belong to the same species (22).
Principal components analysis (PCA) was performed with UNIFRAC (23) by using clustering at 97% sequence identity. The
DNA sequences were deposited in the MG-RAST server database
(http://metagenomics.anl.gov/) with access numbers 4475188.3
4475213.3, 4475215.3, and 4475220.34475235.3 under the project
name Milk Microbiome.

Milk bacterial diversity across lactation time

After quality filtering and length trimming, 119,652 16S rRNA
sequences were analyzed, with an average number of taxonomically
assigned, high-quality sequences of 2623 per sample. The taxonomic assignment of the sequences showed that the composition of
human breast milk is dominated by bacilli, which account for .76%
of the total number of sequences obtained (Figure 1A). In colostrum, the most common genera were Weisella and Leuconostoc
(both lactic acid bacteria from the order Lactobacillales) followed
by Staphylococcus, Streptococcus, and Lactococcus (Figure 1B).
Although lactic acid bacteria genera were still among the most
abundant in 1- and 6-mo milk samples, typical inhabitants of the
oral cavity such as Veillonella, Leptotrichia, and Prevotella or
members of the TM7 phylum (24) increased significantly.
Milk from colostrum compared with 6-mo milk samples
showed different patterns of bacterial diversity. The rarefaction
curves indicated .1000 OTUs when sequences were clustered at
97% sequence identity (the consensus value for determining
species boundaries), .800 OTUs when clustered at 95%, and
.400 OTUs when clustered at 90% (Figure 2). With the sequence
length (400-bp average length) and potential pyrosequencing
errors taken into consideration (25), the high number of OTUs
obtained, even at conservative clustering values, points to several
hundred species in human breast milk, with colostrum having
higher diversity than transition and mature milk.
Milk microbiome and maternal obesity
Maternal BMI influenced the milk microbiome composition. A
PCAwas performed for milk samples from mothers who had delivered
their infants vaginally to eliminate the effect of mode of delivery.
Clustering of samples in obese women indicated a more homogenous
bacterial composition in comparison with the milk samples of normalweight women, which were more scattered (Figure 3A), which
suggests that bacterial composition may have been influenced by
body weight. Further PCAs clearly separated samples from mothers
who gained excessive weight during pregnancy from those of
mothers with normal weight increase (Figure 3B).
The qPCR data corroborated the differences in the proportion of
specific taxonomic groups between obese and normal-weight
mothers (see Table S1a under Supplemental data in the online
issue) and those with excessive and normal weight gain over
pregnancy (see Table S1b under Supplemental data in the online
issue). Higher maternal BMI was related to higher numbers of
Lactobacillus in colostrum (r = 0.600, P = 0.026). Similarly, higher
numbers of Staphylococcus (r = 0.560, P = 0.038) and lower
numbers of Bifidobacterium in breast milk 6 mo postpartum (r =
20.651, P = 0.012) were related to higher maternal BMI. With the
use of mixed models to analyze the effect of BMI on breast-milk
microbiota composition during lactation, we detected higher total
bacteria counts (ratio: 0.34; 95% CI: 0.08, 0.60; P = 0.011),
Staphylococcus (ratio: 0.62; 95% CI: 0.30, 0.93; P = 0.0001), and
Lactobacillus (ratio: 0.52; 95% CI: 0.02, 2.02; P = 0.038) and
lower Bifidobacterium numbers (ratio: 20.48; 95% CI: 20.78,
0.18; P = 0.002) in obese compared with normal-weight women
over the first 6 mo of breastfeeding. Excessive weight gain during
pregnancy was also associated with higher amounts of the genus
Staphylococcus and Staphylococcus aureus (P = 0.09 and 0.03

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

The first 500 base pairs (bp) of the 16S rRNA genes were
amplified with the universal eubacterial primers 27F and 533R by
using high-fidelity AB-Gene DNA polymerase (Thermo Scientific) with an annealing temperature of 528 C and 20 cycles to
minimize PCR biases (19). A secondary amplification was
performed by using the purified PCR product as a template, in
which the universal primers were modified to contain the pyrosequencing adaptors A and B and an 8-bp barcode specific
to each sample, following the method used in Cole et al (20).
Barcodes differed from each other in $3 nucleotides to avoid
errors in sample assignments. Three secondary PCRs were
performed per sample; the PCR products were pooled before
purification by using an Ultrapure PCR purification kit (Roche).
The final DNA per sample was measured by picogreen fluorescence in a Modulus 9200 fluorimeter from Turner Biosystems. PCR products were pyrosequenced from the forward
primer end only by using a GS-FLX sequencer with Titanium
chemistry (Roche) at the Center for Public Health Research in
Valencia, Spain. One-eighth of a plate was used for each pool of
20 samples.

RESULTS

FACTORS INFLUENCING HUMAN MILK MICROBIOME

547

respectively) in 1-mo samples, as well as higher amounts of Lactobacillus (P = 0.03) and lower amounts of Bifidobacterium (P =
0.03) in 6-mo samples (see Table S1b under Supplemental data
in the online issue).

Milk microbiome and mode of delivery

A striking difference in bacterial taxonomic composition was
found between mothers who delivered their infants vaginally and
those who delivered by cesarean section. Mothers who gave birth
by elective cesarean delivery showed a significant compositional
shift, with decreased amounts of Leuconostocaceae and increased amounts of Carnobacteriaceae, among others, compared with those who delivered vaginally (Figure 1, C and D).
This difference was already present in the colostrum and was
maintained in breast milk at 1 and 6 mo. qPCR analysis of the
samples with the use of universal primers indicated that the total
bacterial load in the colostrum was not different between elective and nonelective cesarean-section deliveries (see Table S1c

under Supplemental data in the online issue). Interestingly,

milk samples from mothers who gave birth by nonelective cesarean delivery displayed a composition more similar to milk
from mothers who gave birth by vaginal delivery than to milk
from mothers who gave birth by elective cesarean delivery
(Figure 1, C and D).
Clustering of the different samples according to their composition showed that the milk microbiota was not associated with
any mucosal, fecal, or skin samples. PCA showed that milk
microbiota with a skewed bacterial composition as a result of the
influence of mode of delivery clustered with microbes from
the oral and vaginal mucosa (Figure 4), which suggests that the
bacterial transfer mechanism could be modified.

DISCUSSION

The results of the current study provide new insight into the
composition of microbiota in breast milk. In addition to nutritional
support, breast milk provides bioactive constituents that both

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

FIGURE 1. Bacterial taxonomic composition of human breast milk. The graphs show the proportion of bacterial families (left) and genera (right) as
inferred by polymerase chain reaction amplification and pyrosequencing of the 16S rRNA. A and B: bacterial composition of the milk microbiome through
time (colostrum, 1 and 6 mo after delivery) in normal-weight (n = 8) and obese (n = 10) mothers. C (colostrum samples) and D (milk samples during
breastfeeding period): bacterial composition of milk from mothers who delivered vaginally (n = 9) and by nonelective (n = 3) and elective (n = 6) cesarean
section. Col-NW, colostrum samples from normal-weight mothers; Col-OW, colostrum samples from obese mothers; C-section, cesarean section; 1M-NW,
milk samples at 1 mo of breastfeeding in normal-weight mothers; 1M-OW, milk samples at 1 mo of breastfeeding in obese mothers; 6M-NW, milk samples at
6 mo of breastfeeding in normal-weight mothers; 6M-OW, milk samples at 6 mo of breastfeeding in obese mothers.

548

CABRERA-RUBIO ET AL

directly and indirectly enhance mucosal barrier function and

shape immune development (5, 6). Among these components, our
work shows an extraordinarily diverse array of microorganisms
whose potential functional role should be evaluated. The data also
indicate significant changes in milk microbiome composition and
diversity through lactation in a manner similar to what has been
reported for the protein and fat content of breast milk (7). The
higher amount of diversity and the different bacterial composition
found in our data sets when compared with the other available
data (see Figure S1 under Supplemental data in the online
issue) could be a result of geographic and environmental differences in the mothers and secondly of differences in the experimental protocol [eg, the use of nondegenerate primers, a
high annealing temperature, and a large number of cycles in the
previous study (17)], which taken together may reduce the
identified diversity.
In our samples, the first and second components in PCA clearly
separated the colostrum samples from the 1- and 6-mo breastmilk samples (Figure 3, A and B, respectively), indicating that the
initial colostrum is compositionally distinct from later milk
samples. Although pyrosequencing of the 16S rRNA gene allows
an extraordinary degree of detail in the analysis of microbial
diversity, this approach has limitations such as PCR amplification
biases (19). This is well known in the case of some bacterial
groups, which are not easily amplified by universal primers,
including bifidobacteria, whose presence is a hallmark of the gut
microbiota in healthy breastfed infants (9). Hence, qPCR experiments with the use of primers to amplify specific taxonomic
groups were also performed, including taxa such as Bifidobacterium spp., whose high G+C content disfavors amplification
under standard conditions.
Our results show that maternal BMI and weight gain during
pregnancy have an impact on breast-milk microbiome taxonomic
composition and diversity. Colostrum and 1-mo milk samples
from obese mothers showed a lower diversity than did those from
normal-weight mothers, although the difference disappeared in
milk samples taken at 6 mo (Figure 2A). The interlink between

the gut microbiota and obesity has been extended to high prepregnancy body weight and excessive weight gain during
pregnancy, which act as an inoculum for the gut microbiota in
infants (11, 12). The qPCR data also showed significant differences in the presence of specific bacterial genera depending on
BMI and weight gain during pregnancy, corroborating the conception that obesity influences the milk microbiome. Nevertheless, it must be emphasized that the sample size (18 mothers
were sampled in the present study) limits the interpretation of the
results, and the suggested trends would need to be confirmed by
larger-population studies. Given that milk bacteria are among the
first bacterial cells transferred to the infants orogastrointestinal
tract, a skewed composition in the milk could be a factor contributing to acquire shifts in bacterial composition from mothers
to infants. Thus, our data may indicate an additional mechanism
explaining the heightened obesity risk in infants of obese and
overweight mothers.
Industrialized countries worldwide are experiencing a progressive increase in immunologic and metabolic diseases, and
the velocity of this increase is particularly conspicuous in
children. The mode of delivery is known to influence the
neonatal gut microbiota composition (15, 16). Our data further
show that breast milk from mothers who had given birth by
elective cesarean delivery showed a significant compositional
shift compared with those who had a nonelective cesarean
section and those who delivered vaginally (Figure 1, C and D;
Figure 4). The difference was already present in the colostrum
and was maintained in breast milk at 1 and 6 mo, indicating that
the shift in bacterial composition had a long-term effect. The
fact that the milk microbiome of mothers who gave birth by
nonelective cesarean section had a normal microbial composition that was comparable to that of breast milk from mothers
who delivered vaginally suggests that physiologic (eg, hormonal) changes produced in the mother during the labor process
may influence the composition of the bacterial community.
Given the increased risk of diseases such as allergic rhinitis,
asthma, and celiac disease reported in children born by cesarean

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

FIGURE 2. Diversity of the human milk microbiome. Graphs show rarefaction curves relating the sequencing effect compared with an estimate of the
number of bacterial species, as inferred by the number of OTUs. An OTU was a cluster of 16S rRNA sequences that were .95% identical, a conservative
estimate for the boundary between species, established at 97% for full-length 16S gene sequences. Panel A shows rarefaction curves for milk samples through
time (colostrum, 1 and 6 mo after delivery) from normal-weight and obese mothers. Panel B shows rarefaction curves for the 3 delivery modes studied. ColNW, colostrum samples from normal-weight mothers; Col-OW, colostrum samples from obese mothers; C-section, cesarean section; OTUs, operational
taxonomic units; 1M-NW, milk samples at 1 mo of breastfeeding in normal-weight mothers; 1M-OW, milk samples at 1 mo of breastfeeding in obese mothers;
6M-NW, milk samples at 6 mo of breastfeeding in normal-weight mothers; 6M-OW, milk samples at 6 mo of breastfeeding in obese mothers.

FACTORS INFLUENCING HUMAN MILK MICROBIOME

549

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

FIGURE 3. PCoA on the basis of the bacterial composition of human breast milk. A and B: Distribution of colostrum and 1- and 6-mo samples from obese and
normal-weight mothers. C and D: Distribution of colostrum and 1- and 6-mo samples from mothers with normal weight gain and excessive weight gain during
pregnancy. All samples were collected from mothers who gave birth by vaginal delivery (n = 9). Col-EWG, colostrum samples from excessive-weight-gain
mothers; Col-NW, colostrum samples from normal-weight mothers; Col-NWG, colostrum samples from normal-weight-gain mothers; Col-OW, colostrum samples
from obese mothers; P, component of the principal coordinates analysis; PCoA, principal coordinates analysis; 1M-EWG, milk samples at 1 mo of breastfeeding in
excessive-weight-gain mothers; 1M-NW, milk samples at 1 mo of breastfeeding in normal-weight mothers; 1M-NWG, milk samples at 1 mo of breastfeeding in
normal-weight-gain mothers; 1M-OW, milk samples at 1 mo of breastfeeding in obese mothers; 6M-EWG, milk samples at 6 mo of breastfeeding in excessiveweight-gain mothers; 6M-NW, milk samples at 6 mo of breastfeeding in normal-weight mothers; 6M-NWG, milk samples at 6 mo of breastfeeding in normalweight-gain mothers; 6M-OW, milk samples at 6 mo of breastfeeding in obese mothers.

delivery (28, 29), further evaluation of the possible impact of the

skewed bacterial composition of breast milk needs to be performed.
It is a widely held belief that sources of colonization of the
bacteria in breast milk include the skin and oral cavity of newborns
exposed to the mothers vaginal and intestinal microbiota during

childbirth (15). Therefore, to shed light on the potential origin of

the milk microbiome, its microbiota was compared with the
available data for female skin, vaginal, oral cavity, fecal, and gut
mucosal microbiota, including a total of 500,000 16S rRNA
sequences. Our results suggest that the milk microbiome is

550

CABRERA-RUBIO ET AL

compositionally distinct from any other human niche and that it is

not a simple contaminant from the skin (Figure 4), although part
of the differences could be a result of the different sampling
procedures. Clustering of the different samples according to their
composition showed that the milk microbiota was not associated
with any mucosal or fecal samples (data not shown). Thus, the
milk microbiome does not appear to be a subset of any specific
human niche. In addition, the suggested transfer of bacteria
present in the human gut to reach the mammary gland through an
endogenous route (the so-called entero-mammary pathway) via
the mucosal associated lymphoid system (30, 31) remains unclear.
It has been reported that a bacterial peptidoglycan originating
from gut microbes could translocate into the circulation and serve
as a molecular mediator responsible for the remote systemic
priming of neutrophils in the bone marrow (32). Another theory
suggested is endocytosis, which could be supported by our observed differences between elective and nonelective cesarean
sections and the stress of labor and the delivery process affecting
gut permeability. There should be a selective process by which
only specific taxa are transferred or able to survive to form the
characteristic, lactic aciddominated microbiota of the milk. This
theory is further supported by reports that the enteric nerve system
affects gut sampling of bacteria and their transfer to Peyer patches
(33, 34). Another source of colonization of these bacteria could be
the skin and oral cavity of newborns exposed to the mothers
vaginal and intestinal microbiota during childbirth (15). We found
an increase in typical oral inhabitants in 1- and 6-mo-old milk,
suggesting that during breastfeeding and nursing, bacteria from
the infants mouth could colonize the milk ducts and areola.
However, recent molecular studies have shown that milk secreted

We thank A Durban and collaborators for sharing the 16S rRNA sequences
from gut biopsies and MA McGuire and collaborators for sharing their sequence data from milk samples in American donors.
The authors responsibilities were as followsMCC and AM: designed
the study and wrote the manuscript; RC-R, MCC, KL, SS, EI, and AM:
carried out the research; and RC-R, MCC, and AM: analyzed the data. All
of the authors revised and approved the manuscript. None of the authors had
any conflicts of interest.

REFERENCES
1. Oftedal OT. The mammary gland and its origin during synapsid evolution. J Mammary Gland Biol Neoplasia 2002;7:22552.
2. Petherick A. Development: mothers milk: a rich opportunity. Nature
2010;468:S57.
3. Walker A. Breast milk as the gold standard for protective nutrients.
J Pediatr 2010;156:S37.
4. LeBouder E, Rey-Nores JE, Raby AC, Affolter M, Vidal K, Thornton
CA, Labeta MO. Modulation of neonatal microbial recognition: TLRmediated innate immune responses are specifically and differentially
modulated by human milk. J Immunol 2006;176:374252.
5. Brandtzaeg P. ABC of mucosal immunology. Nestle Nutr Workshop
Ser Pediatr Program 2009;64:2338.
6. Stockinger S, Hornef MW, Chassin C. Establishment of intestinal homeostasis during the neonatal period. Cell Mol Life Sci 2011;68:
3699712.
7. Hoppu U, Isolauri E, Laakso P, Matomaki J, Laitinen K. Probiotics and
dietary counselling targeting maternal dietary fat intake modifies breast
milk fatty acids and cytokines. Eur J Nutr 2012;51:2119.
8. Gronlund MM, Gueimonde M, Laitinen K, Kociubinski G, Gronroos T,
Salminen S, Isolauri E. Maternal breast-milk and intestinal bifidobacteria guide the compositional development of the Bifidobacterium
microbiota in infants at risk of allergic disease. Clin Exp Allergy 2007;
37:176472.
9. Martn R, Jimenez E, Heilig H, Fernandez L, Marn ML, Zoetendal
EG, Rodrguez JM. Isolation of bifidobacteria from breast milk
and assessment of the bifidobacterial population by PCR-denaturing
gradient gel electrophoresis and quantitative real-time PCR. Appl
Environ Microbiol 2009;75:9659.
10. Roger LC, Costabile A, Holland DT, Hoyles L, McCartney AL. Examination
of faecal Bifidobacterium populations in breast- and formula-fed infants
during the first 18 months of life. Microbiology 2010;156:332941.
11. Aaltonen J, Ojala T, Laitinen K, Poussa T, Ozanne S, Isolauri E. Impact
of maternal diet during pregnancy and breastfeeding on infant metabolic

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

FIGURE 4. Relatedness between milk bacteria and the rest of the human
microbiome. The graph shows a principal components analysis of the
bacterial composition of milk from colostrum and at 1 and 6 mo compared
with available adult female sequences from the same 16S gene region
(hypervariable regions V1 and V2) from skin, vagina, feces, gut mucosa,
and oral epithelium. Milk samples from the 3 time points clustered
together and separately from the rest, except for samples collected from
mothers who gave birth by elective and nonelective cesarean section,
which overlap with oral samples. Data from the human microbiome are
from fecal samples (26); from skin, vagina, and oral samples (15); and from
Durban et al (27). C-section, cesarean section; Milk1, milk at 1 mo; Milk6,
milk at 6 mo; PC, principal components; PC1, first component of the principal
components analysis; PC2, second component of the principal components
analysis; PC3, third component of the principal components analysis.

by the mammary glands during the weeks previous to labor (and

thus not submitted to any kind of infant contact) contains bacterial
species similar to those isolated from fresh milk obtained after
labor (35). Dendritic cells have been described to penetrate the
intestinal epithelium to take up commensal bacteria from the gut
lumen (36), to reach the systemic circulation, and to retain even
live bacteria for several days (37); recently, the transfer of
intestinal bacteria to the mammary glands within dendritic
cells has been proposed (28, 38).
Prenatal and postnatal microbial exposures have profound
effects on the microbial colonization of the intestine and maturation of the naive immune system. Given that the bacteria
present in breast milk are among the very first microbes that enter
the human body and given the vital role of bacteria in the infants
physiology and development of the immune system, our data
emphasize the necessity to understand the biological role that
the breast milk microbiome could potentially play for human
health. If the bacterial composition of human breast milk has
coevolved to maximize the infants metabolic efficiency and to
optimally stimulate the immune system, a skewed microbial
milk composition might have important consequences for the
infants health, and those potential consequences would need to
be evaluated for future recommendations in child nutrition.

FACTORS INFLUENCING HUMAN MILK MICROBIOME

12.
13.
14.

15.

16.

18.
19.

20.

21.

22.
23.

24.

25. Quince C, Lanzen A, Curtis TP, Davenport RJ, Hall N, Head IM, Read
LF, Sloan WT. Accurate determination of microbial diversity from 454
pyrosequencing data. Nat Methods 2009;6:63941.
26. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley
RE, Sogin ML, Jones WJ, Roe BA, Affourtit JP, et al. A core gut
microbiome in obese and lean twins. Nature 2009;457:4804.
27. Durban A, Abellan JJ, Jimenez-Hernandez N, Salgado P, Ponce M,
Ponce J, Garrigues V, Latorre A, Moya A. Structural alterations of
faecal and mucosa-associated bacterial communities in irritable bowel
syndrome. Environ Microbiol Rep 2012;4:2427.
28. Thavagnanam S, Fleming J, Bromley A, Shields MD, Cardwell CR. A
meta-analysis of the association between Caesarean section and
childhood asthma. Clin Exp Allergy 2008;38:62933.
29. Decker E, Engelmann G, Findeisen A, Gerner P, Laass M, Ney D,
Posovszky C, Hoy L, Hornef MW. Cesarean delivery is associated with
celiac disease but not inflammatory bowel disease in children. Pediatrics 2010;125:e143340.
30. Perez PF, Dore J, Leclerc M, Levenez F, Benyacoub J, Serrant P, SeguraRoggero I, Schiffrin EJ, Donnet-Hughes A. Bacterial imprinting of the
neonatal immune system: lessons from maternal cells? Pediatrics 2007;
119:e72432.
31. Donnet-Hughes A, Perez PF, Dore J, Leclerc M, Levenez F, Benyacoub
J, Serrant P, Segura-Roggero I, Schiffrin EJ. Potential role of the intestinal microbiota of the mother in neonatal immune education. Proc
Nutr Soc 2010;69:40715.
32. Philpott DJ, Girardin SE. Gut microbes extend reach to systemic innate
immunity. Nat Med 2010;16:1601.
33. Green BT, Lyte M, Kulkarni-Narla A, Brown DR. Neuromodulation of
enteropathogen internalization in Peyers patches from porcine jejunum. J Neuroimmunol 2003;141:7482.
34. Nickel T, Emslander I, Sisic Z, David R, Schmaderer C, Marx N,
Schmidt-Trucksass A, Hoster E, Halle M, Weis M, et al. Modulation of
dendritic cells and toll-like receptors by marathon running. Eur J Appl
Physiol 2012;112:1699708.
35. Martn R, Langa S, Reviriego C, Jimenez E, Marn ML, Olivares M,
Boza J, Jimenez J, Fernandez L, Xaus J, et al. The commensal microflora of human milk: new perspectives for food bacteriotherapy and
probiotics. Trends Food Sci Technol 2004;15:1217.
36. Rescigno M, Rotta G, Valzasina B, Ricciardi-Castagnoli P. Dendritic
cells shuttle microbes across gut epithelial monolayers. Immunobiology 2001;204:57281.
37. Macpherson AJ, Uhr T. Compartmentalization of the mucosal immune
responses to commensal intestinal bacteria. Ann N Y Acad Sci 2004;
1029:3643.
38. Qutaishat SS, Stemper ME, Spencer SK, Borchardt MA, Opitz JC,
Monson TA, Anderson JL, Ellingson JL. Transmission of Salmonella
enterica serotype typhimurium DT104 to infants through mothers
breast milk. Pediatrics 2003;111:14426.

Downloaded from ajcn.nutrition.org at UKK INSTITUUTTI KIRJASTO on September 4, 2012

17.

programming: a prospective randomized controlled study. Eur J Clin

Nutr 2011;65:109.
Kalliomaki M, Collado MC, Salminen S, Isolauri E. Early differences
in faecal microbiota composition in children may predict later weight
gain? Am J Clin Nutr 2008;87:5348.
Collado MC, Isolauri E, Laitinen K, Salminen S. Distinct composition
of gut microbiota during pregnancy in overweight and normal-weight
women. Am J Clin Nutr 2008;88:8949.
Collado MC, Isolauri E, Laitinen K, Salminen S. Effect of mothers
weight on infants microbiota acquisition, composition, and activity
during early infancy: a prospective follow-up study initiated in early
pregnancy. Am J Clin Nutr 2010;92:102330.
Dominguez-Bello MG, Costello EK, Contreras M, Magris M, Hidalgo
G, Fierer N, Knight R. Delivery mode shapes the acquisition and
structure of the initial microbiota across multiple body habitats in
newborns. Proc Natl Acad Sci USA 2010;107:119715.
Huurre A, Kalliomaki M, Rautava S, Rinne M, Salminen S, Isolauri E.
Mode of deliveryeffects on gut microbiota and humoral immunity.
Neonatology 2008;93:23640.
Hunt KM, Foster JA, Forney LJ, Schutte UM, Beck DL, Abdo Z, Fox
LK, Williams JE, McGuire MK, McGuire MA. Characterization of the
diversity and temporal stability of bacterial communities in human
milk. PLoS ONE 2011;6:e21313.
Laitinen K, Poussa T, Isolauri E. Probiotics and dietary counselling
contribute to glucose regulation during and after pregnancy: a randomised controlled trial. Br J Nutr 2009;101:167987.
McKenna P, Hoffmann C, Minkah N, Aye PP, Lackner A, Liu Z,
Lozupone CA, Hamady M, Knight R, Bushman FD. The macaque gut
microbiome in health, lentiviral infection, and chronic enterocolitis.
PLoS Pathog 2008;4:e20.
Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-SyedMohideen AS, McGarrell DM, Marsh T, Garrity GM, et al. The Ribosomal Database Project: improved alignments and new tools for
rRNA analysis. Nucleic Acids Res 2009;37:D1415.
Yarza P, Richter M, Peplies J, Euzeby J, Amann R, Schleifer KH,
Ludwig W, Glockner FO, Rossello-Mora R. The All-Species Living
Tree project: a 16S rRNA-based phylogenetic tree of all sequenced
type strains. Syst Appl Microbiol 2008;31:24150.
Lozupone C, Hamady M, Knight R. UniFracan online tool for
comparing microbial community diversity in a phylogenetic context.
BMC Bioinformatics 2006;7:371.
Sipos R, Szekely AJ, Palatinszky M, Revesz S, Marialigeti K, Nikolausz
M. Effect of primer mismatch, annealing temperature and PCR cycle
number on 16S rRNA gene-targetting bacterial community analysis.
FEMS Microbiol Ecol 2007;60:34150.
Belda-Ferre P, Alcaraz LD, Cabrera-Rubio R, Romero H, Simon-Soro
A, Pignatelli M, Mira A. The oral metagenome in health and disease.
ISME J 2012;6:4656.

551